Lipids and lipoproteins have been associated with breast cancer risk; however, published results have been inconsistent. To clarify these associations, we measured fasting lipids in women undergoing breast biopsies. A case-control study examined the association of fasting levels of lipids with histologically defined breast cancer risk. Four groups of premenopausal women were assembled on the basis of histological appearance of breast tissue: 1) no epithelial proliferation (n = 102), 2) proliferation without atypia (n = 53), 3) atypical hyperplasia or carcinoma in situ (n = 53), and 4) node-negative invasive cancer (n = 102). A postoperative fasting blood specimen was analyzed for cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglycerides. Demographics, risk factors, diet, physical activity, fasting weight, and skin-fold thickness were measured. Triglyceride levels were significantly higher in women with node-negative invasive cancer (0.94 +/- 1.04 mg/ml) than in those with no epithelial proliferation (0.83 +/- 1.04 mg/ml, p = 0.03). This association persisted after adjustment for age, body size, lipids, reproductive and familial risk factors, and previous benign breast problems (p < 0.01), in keeping with an independent association of elevated triglycerides with breast cancer risk.